SuperAPI AI Directory
Partner中文AI newsFavorites
Skills01-intake

01-intake

General & other

View repository

Updated 10-16 05:00

About this skill

01 · Intake — Company Snapshot

The first step of the GTM flow. Assemble from the public website of the agency a structured snapshot on which all subsequent steps are built. This is cold audit — no client data, only what is visible from the outside.

Input: the site URL ($ARGUMENTS).

Reuse

  • First call the skill website-scraper to collect pages (home, services, cases, about, pricing, blog, careers, contact).
  • Then call deep-company-analyser to extract pain points, customer language, proof, and positioning signals.
  • If these skills are absent — do the same via WebFetch/WebSearch on the key pages of the site.
  • SPA fallback: if the site is client-rendered (WebFetch/rag-web-browser return only <title>), use Apify apify/website-content-crawler (crawlerType: playwright, dynamicContentWaitSecs).
  • LinkedIn team (Apify): for section 11 find the company LinkedIn (web search "<назва>" site:linkedin.com), then: scrapemint/linkedin-company-employees-scraper (free overview: headcount + role distribution + featured) and harvestapi/linkedin-company-employees (roster; filter functionIds 25=Sales, 15=Marketing, 4=Business Development). Founder — web search "<назва>" founder/CEO + LinkedIn.

What to collect (Company Snapshot)

  1. Identification: name, domain, year/“since”, geo signals, site languages.
  2. Services: list of services as the site calls them (verbatim) + whether there is a focus or “everything for everyone”.
  3. Positioning (as is): headline/hero, whom it is addressed to, stated uniqueness, tagline. + Positioning scoring (cold). Evaluate ONLY what a cold visitor sees within 5 seconds on positioning surfaces (hero/H1/tagline/above-the-fold of the home page). Each item 0 / 0.5 / 1 + verbatim proof or “none”:
    • Positioning — X out of 5: (1) ICP named; (2) industry/category named; (3) the problem you solve is clear; (4) use case is clear; (5) offer is clear.
    • Value Proposition — X out of 4 (Hormozi value equation): (1) the audience’s Dream Outcome; (2) proofs (perceived likelihood); (3) Time Delay to get the result; (4) Effort & Sacrifice (what is required from the client / how painless it is). This is cold scoring “as is” — raw material for 06-positioning (there the enemy angle and Z1.4 self-score), not the final verdict.
  4. Target hints: industries/client size/project types that are mentioned.
  5. Proof: cases (how many, whether with $-results, named/anonymous), client logos, testimonials, awards, Clutch/ratings.
  6. Pricing signals: pricing/tiers/"from $", "contact us", productized vs bespoke.
  7. Team: size (from about/careers/LinkedIn hints), visible roles (founder, sales, marketing, AM), open BizDev role vacancies.
  8. Presence channels: company LinkedIn (active/empty), founder's personal LinkedIn, blog/SEO, lead magnets, Upwork/Clutch profiles.
  9. Technology signals: stack (Wappalyzer-style, if available), site quality (landing / full / service pages / ROI calculators).
  10. Client language (verbatim): 5–15 phrases from the site with which the agency describes problems/benefits — raw material for VP and copy in the next steps.
  11. LinkedIn: team and bizdev (Apify). A separate section — the real team composition, not from the site:
    • Headcount (current) + growth trend 6m/1y/2y + median tenure — the main dynamics signals. Sources by priority: (1) li_at cookie actor (LinkedIn session with userConfig.linkedin_li_at): pulls the full roster + staffCount, and a Sales-Navigator-aware actor — also the 6m/1y/2y trend + median tenure; (2) fallback — LinkedIn Premium "Total employee count" inserted by the consultant (screenshot/number); (3) anonymous actors often hit a login wall (name: "Join LinkedIn" / null) → trend unavailable, mark "li_at / Premium required".
    • Role distribution (scrapemint) + roster (harvestapi).
    • BizDev functions: how many Sales / Marketing / SDR / BDM / AM (functions 25/15/4). 0 visible bizdev roles = founder-led, no engine (a strong proxy for Stage 0-1).
    • Founder(и): name, background, whether it matches the site's claims (e.g. "20+ years" = the founder's personal experience or a team entity?).
    • Entity age: the earliest tenure/"since" in the profiles → how new the brand is.
    • Tech signals from profiles: stack in role descriptions (often more accurate than the site — narrows the ICP/offer).
    • ⚠️ Public LinkedIn undercounts unlinked profiles — this is a headcount floor, not an exact number.
  12. Traffic and SEO (Ahrefs MCP + SimilarWeb). A mandatory section of every snapshot:
    • Ahrefs MCP (main source): DR + backlinks stats, site-explorer-metrics-history (organic traffic for 24 months, monthly, mode=subdomains), top pages by traffic, top organic keywords. Before first use, check the units via subscription-info-limits-and-usage (free) — The Trial account has 0 API units, then mark “Ahrefs unavailable, collect after upgrade” and do not burn calls.
    • Semrush MCP (server semrush, https://mcp.semrush.com/v1/mcp, user scope, OAuth): as of 2026-06-10 Viktor's plan does not include MCP access — the tools respond with a refusal about the plan. Try it as the first call; if refused — note it and move on.
    • SimilarWeb (free, main working route): (а) lite-endpoint https://data.similarweb.com/api/v1/data?domain=<домен> via WebFetch — visits/month (3 months of history), bounce/pages/duration, geo, rankings; (b) Apify actor tri_angle/similarweb-scraper (permissions confirmed, ~$0.06/domain) — adds top organic queries, similar competitors, geo-share changes, technographics. For sites <50K/month the figures are approximate — note this. epctex/similarweb-scraper — paid subscription $15/month, do NOT use.
    • Interpretation is mandatory: traffic growing/flat/declining; whether it converts or not (lead magnets? CTA?); whether the geo matches the target market. This is raw material for the SMF gap L and 11-channels-plan.

Stage proxy signals

As a separate block, mark which of the found signs indicate stage 0/1/2/3/4 (see ${CLAUDE_PLUGIN_ROOT}/reference/agency-stages.md, section 1 — cold-audit proxy signals). This is raw material for 02-stage-diagnostic, do not declare the stage here yet.

Output (Notion)

Create/update the subpage «01 · Company Snapshot» on the client page (see ${CLAUDE_PLUGIN_ROOT}/reference/notion-schema.md). Structure: sections 1–12 above + block “Stage proxy signals” + block “Sources / assumptions” (which pages were read, where the hypothesis is). In Ukrainian; service names and verbatim phrases — in the original language.

Definition of Done

  • All 12 sections filled in (where there is no data — explicitly “not found”, do not invent).
  • Section 3 contains scoring: Positioning X/5 + Value Proposition X/4, each item with a verbatim proof.
  • Section 12: traffic overview collected (Ahrefs or SimilarWeb fallback); if Ahrefs without units — this is explicitly marked in the section and in "Sources / assumptions".
  • ≥ 5 verbatim phrases in the client's language collected.
  • Section 11: company LinkedIn checked (found/not found); business dev roles counted; founder identified or explicitly "not publicly found".
  • Block of stage proxy signals ready for step 2 (taking into account team composition).
  • Sub-page in Notion created; a short summary returned to the orchestrator.